Now, we finally have a Civitai SD webui extension!!
Update:
v1.10.2 is on github, which support SD webui v1.8.x.
Some models request login to civitai to download. So you need to set a civitai API Key on Civitai, then fill that API Key into this extension's setting page. Check document on github (Civitai API Key Section) for detail.
Civitai Helper
Stable Diffusion Webui Extension for Civitai, to help you handle models much more easily.
The official SD extension for civitai takes months for developing and still has no good output. So, I developed this Unofficial one.
Github project:
https://github.com/butaixianran/Stable-Diffusion-Webui-Civitai-Helper
(Github page has better document)
Feature
Scan all models to download model information and preview images from Civitai.
Link local model to a civitai model by a civitai url
Download a model(with info+preview) by Civitai Url into SD's model folder or subfolder.
Downloading can resume at break-point.
Checking all your local model's new version from Civitai
Download a new version directly into SD model folder (with info+preview)
Modified Built-in "Extra Network" cards, to add the following buttons on each card:
🌐: Open this model's Civitai url in a new tab
💡: Add this model's trigger words to prompt
🏷: Use this model's preview image's prompt
Install
Everytime you install or update this extension, you need to shutdown SD Webui and Relaunch it. Just "Reload UI" won't work.
How to use
First of all, Update Your SD Webui to latest version!
This extension need to get extra network's cards id. Which is added since 2023-02-06. If your SD webui is an earlier version, you need to update it!
After install, Go to extension tab "Civitai Helper". There is a button called "Scan Model".
Click it, extension will scan all your models to generate SHA256 hash, and use this hash, to get model information and preview images from civitai.
After scanning finished,
Go to SD webui's build-in "Extra Network" tab, and click "Refresh Civitai Helper" button, to show all addtional buttons.
Everytime extra network tab refreshed, it will remove all additional buttons of this extension. You need to click Refresh Civitai Helper button to bring them back.
Move your mouse on to the Top of a model card. It will show 3 additional icon buttons:
🌐: Open this model's Civitai url in a new tab
💡: Add this model's trigger words to prompt
🏷: Use this model's preview image's prompt
Description
v1.5.7
Fix Localization issue for 4 addtional buttons on cards
FAQ
Comments (167)
OMG! Can't wait to test it!
Can it download civit models directly in webui when provided with a link or hash?
Yes, i want civitai catalog that download loras, embeddings, hypervenerworks and checkpoints with preview to the correct folders too!
That's not a good idea. When it is downloading, your SD webui gonna be stoped at there, you can not do anything until the downloading is finish.
@butaixianran Then the extension is badly programmed. It should be able to do things in the background.
I wait here for the reports for using feedback.
great work ! thank you indeed!
WOOOOOOOOOOOOW!!!!!! Thanks!!!
I'm working also with something to work better with civitai.
I'll show something asap.
Good Job!
Experienced insane slowdowns during scanning. Surely it doesn't need to read entirety of models to scan the hash? Isn't it at the very beginning of file?
I sure hope it skips files it already scanned when I next have to use this...
first time run it will creat a empty hash to it's own base , second time will jump these hash already have, i still test have not finish yet
That is how hash works, read the whole file piece by piece. If you used any cloud driver, you'll understand. Of cause it won't scan the same file twice.
@butaixianran Yeah I'm almost finished now. I think it could be a good idea to add scanning button separately for each embeddings/hypernetworks/loras/checkpoints, checkpoints are very painful to generate hash for and might cause heavy slowdowns.
sha256 must read the full contents of the file
(autov2 is the front 10 hexadecimal digits of sha256)
lol sorry ... i had to laugh real loud.
but yes it does it is the entire point of hashing. :)
它不能在colab上正常工作，我已经正确安装了扩展，但是没有看到预览图，而且一旦使用
Refresh Civitai Helper会产生更多的重复按钮
It does not work properly on colab, I have installed the extension correctly, but do not see the preview image, and once used
The Refresh Civitai Helper generates more duplicate buttons
If you are working on colab, can you see the process it scanning models?
@butaixianran
Civitai Helper: Start scan_model
Civitai Helper: Scanning path: /content/gdrive/MyDrive/sd/stable-diffusion-webui/embeddings
Civitai Helper: Scanning path: /content/gdrive/MyDrive/sd/stable-diffusion-webui/models/hypernetworks
Civitai Helper: Scanning path: /content/gdrive/MyDrive/sd/stable-diffusion-webui/models/Stable-diffusion
Civitai Helper: Generate SHA256 for model: anything-v4.5.safetensors tcmalloc: large alloc 7703281664 bytes == 0x7f021ad94000 @ 0x7f075426e1e7 0x5f97c1 0x649901 0x5c43c6 0x4f327e 0x64e618 0x505163 0x56bbe1 0x5f5ee6 0x56bab6 0x5f5ee6 0x4f327e 0x5ff2a3 0x5c4677 0x5f5878 0x5717f5 0x5f5ee6 0x56bbe1 0x5f5ee6 0x56bbe1 0x5f5ee6 0x50b32c 0x5f52b2 0x6568ec 0x677e08 0x7f075402e609 0x7f0754168133
^C
这是第二次测试：
At second time:
Total progress100% 20/20 [00:05<00:00, 3.59it/s]
Civitai Helper: Start scan_model Civitai Helper: Scanning path: /content/gdrive/MyDrive/sd/stable-diffusion-webui/embeddings Civitai Helper: Scanning path: /content/gdrive/MyDrive/sd/stable-diffusion-webui/models/hypernetworks Civitai Helper: Scanning path: /content/gdrive/MyDrive/sd/stable-diffusion-webui/models/Stable-diffusion Civitai Helper: Generate SHA256 for model: anything-v4.5.safetensors tcmalloc: large alloc 7703281664 bytes == 0x7f9443e8a000 @ 0x7f99adada1e7 0x5f97c1 0x649901 0x5c43c6 0x4f327e 0x64e618 0x505163 0x56bbe1 0x5f5ee6 0x56bab6 0x5f5ee6 0x4f327e 0x5ff2a3 0x5c4677 0x5f5878 0x5717f5 0x5f5ee6 0x56bbe1 0x5f5ee6 0x56bbe1 0x5f5ee6 0x50b32c 0x5f52b2 0x6568ec 0x677e08 0x7f99ad89a609 0x7f99ad9d4133 ^C
你的报错信息我去查了下，是Google Driver和colab比较有名的使用限制。你可以搜索"colab python tcmalloc large alloc" 了解详情。这个是google云服务的已知限制，是由于内存不足，被colab强制退出导致的。高级帐号可以让内存翻倍，如果你不是高级帐号，你可以充值解决。
添加了一个内存优化版本的sha256计算方法，插件界面有个新选项，勾上就行
This is a god send, I was about to write something similar just for trigger word tracking.
I'm getting an error that says that "there is no image in gallery to save as a preview"? Please advise.
So, this model does not has a preview image on civitai? Offer the model url here, so I can check it.
@butaixianran It does that for many of the models on my list. I suspect it has something to do with the fact that the models are not located directly in the Auto 1111 folder, because it reads the Loras that are with no problem. My models are loacted on another SSD. so It may be a directory issue.
@9ball ok, if you are using startup command arguement like ckpt-dir or lora-dir, it does not support that yet. Will be supported in the next version
@butaixianran Yea, everything seems to be working now for me so far. Would it be possible to allow the user to browse the preview images and choose which one they would like to set for each model? Secondly, I don't know how big of an undertaking this would be, but it would be nice to have an extension that allowed you to make annotations/notes on each model. I often find myself spending ram on keeping windows open just so I can remember the prompts and instructions. Either way, thank you very much.
如果模型在子文件夹内，就无效吖
是的，支持子文件夹的还没来得及做，会做的
need a function to manage trigger word
This is awesome. Thanks for sharing!
Is this a different extension from the one that shows on top when you click run model?
[EASY UPDATE within the webui.(for beginners)]
Since a couple updates ago you can put the github link in the extensions tab. This way you can update the extension without downloading again.
Go to Extension tab > Install from URL, and copy the github main page of the extension.
I use a spreadsheet with all of the Names, Categories, Keywords, URLs, and weight values of my TI, LoRA and Hypernetworks. It then concatenates them into the style list format.
Then I use the style list to select my LoRAs with the appropriate weight and keywords.
It would be kind of cool to see how you did that if possible?
thats dope. May have to make myself a database similar to that.
is that a styles.csv sheet ? great idea by the way
awesome!
Wait one task finished then do another
how can it work with lora?please
and what is the momory optimse SHA256? is it better?
V1.2 It doesn't work very well. after I scaned the models, the refresh button and four small icons at the extra network were not displayed
What you got in the console log window? Is there any error msg there?
Also, have you tried to click the "Refresh Civitai Helper" button on extra network toolbar? Since your extra network refreshed after you got new preview images, you need to click that button to refresh this extension too.
@butaixianran Me too，the "Refresh Civitai Helper" button not show too,just tab "Civitai Helper"show
@butaixianran 控制台没有错误提示，安装插件后扫描完毕，可以在extra network正确显示对应模型的图片，这点以前是没有的。但是在extra network框架下没有出现新的“Refresh Civitai Helper”按钮，只有原有的刷新关闭按钮，同样的没有浮于图片上方的几个小按钮。
@4121316314 这个"Refresh Civitai Helper"按钮，和python端完全没有关系，按道理是一定会出现的。而且，这个部分代码根本没有去碰。如果旧版本出现了，新版本按道理应该也出现。
尝试在浏览器中按下F12，打开命令行记录，里面应该有错误提示，把错误提示复制过来。尽量完整。如果这里不方便，就在github项目里开issue。
我在js一侧添加了一些额外的防出错的验证。可以从github更新新版试试看。
@butaixianran 太好了！可以正常运行了，感谢你做了如此细致完美的工作！
离完美相去甚远，这个小插件还有很长的路要走。
在pad之类的移动端操作的话，触屏没法支持操作，四个按钮都不显示😭
1.5版本增加了选项，可以一直显示按钮，支持触屏。
have you tried click the big "Refresh Civitai Helper" button?
OK, I guess I found the reason. You need to update your SD webui to latest version. The extra network cards id is added since 2023-02-06. So, if your SD webui is an earlier version, you need to update it.
how to add custom folder, for Web Ui Colab?, thanks
Please! Add an option, what to do with existing previews: SKIP already existing previews (so don't download new ones from civitai and keep old ones), RENAME them (should be default, old preview still exist but hidden on extra networks tab) or REPLACE (delete existing, current behaviour)
Your own preview image should be named as "model_name.png", while this extension downloaded images are named as "model_name.preview.png".
So, it won't delete your own preview image. And your own preview image("model_name.png") has a higher priority than downloaded image("model_name.preview.png")
This is a build-in feature of extra network. You need to learn how to use it.
@butaixianran ok but will it redownload a preview it already has?
can we have a feature that updates the loras/models to newest version by 1 click?
check new version
@butaixianran you did an amazing job!!
This also can download preview images of the users generated in the commnents?
Love this tool. super useful. One thing I would suggest for next version. since a lot of people tend to keep older/different versions of LoRAs/models due to them being better at certain things. Maybe have the 'check for update' program skip over ones when it notices it has the most recent one in the same folder. Like if it had anything v1, 2 and 3 in the same folder. have it recognize it has the most recent in there and have it skip the update recommendation for v1 and 2. As it is it tells me Ihave 15 models to update when i already have the updates.
done, update your version from github
@butaixianran I updated. doesn't seem to work. Still exact same number of models it said i needed to update before. Like i have version 13, 15, 16, 17, 18 and 19 of one model. I use most of them depending on what I'm trying to do but the updater just tells me i need to update 13, 15, 16, 17 and 18 to 19. Which i have. They are all in the same folder, even.
@derpnope
First of all, have you reloaded UI after updating extension? If you don't, you are still using old version of this extension.
If so, that means, you actually do not have the latest version of this model.
It only ignores a model, when you have the latest version of that model.
Also, it only remind you once, even if you have multiple old versions. So, looks like you are not using the latest version of this extension.
On "Civitai Helper" Tab, at the bottom, there is a version number. What is your numbe there?
@butaixianran 1.4.1
@derpnope One thing I forgot, I think this is the reason. Just Reload UI won't work, you need to shut down SD webui, then re-launch it.
I changed the code library in sub-folder, which won't be reloaded if you just reloading UI
@butaixianran Ah there it goes. that did it. Thanks <3
I don't see the buttons anymore after today's update.
have you tried "Refresh Civitai Helper" button? Have you updated your SD webui like it asked?
WebUI is up to date, and I did refresh the helper.
in your browser, press F12, go to console tab, leave it open, then click "Refresh Civitai Helper" Button, then, what error message you have in the console page?
@butaixianran It worked again after uninstalling and reinstalling the extension. Thanks for your work!
What is this for?
You need to mentioned that what we can do with this at the top.
The giant picture preview lists most/all of the features.
omg this is just too good! ty!
Some preview images are not working, it looks like it's .jfif extensions being used in some of the previews thats causing the problems. for example the lora "Sun and Shadow" preview images are all .jfif
I found just renaming the extension to .png works, no need to convert the .jfif
hi, this extension download all preview images as .png. There is no way it gonna name it as .jfif. No such code in it.
So, you renamed it by yourself? Or, the preview images are not downloaded by this extension?
@butaixianran Hey, No sorry if I was not clear, the extension will only download .png/.jpgs but some of the new Civitai preview images are in a .Jfif Format, so the extension does not see it as a usable image. I had to manually download the image and rename it.
Can you disable "download max size preview" by default? It's not necessary in the first place and takes additional space.
Or at least remember the setting? It keeps resetting back to this value.
at the bottom of extension tab, there is a save setting button, use it
The latest update is pushing my sub folders bar to the side instead of along the top, and is forcing the images to the side and I can't figure out why.
Example here https://i.imgur.com/TlNRNdy.png
It's not a big deal in the example because there's only a few models to pick from, but in categories that have 50+ models then the sub folder / category selector buttons are way the hell down the page, and I have to scroll all the way down when I first open Lora / Checkpoints etc just to change it from All to one of the narrow categories that is relevant
There is nothing changed at client side. What you have here, is a css issue of extra network, which this extension didn't touch it at all.
If you disable this extension and re-start your SD webui, what gonna happen, will this issue still be there?
Also, have you tried to shut down and re-launch SD webui after updating this extension? Like document asked.
If no matter what, this issue is still there, you can try this:
On extra network page, Press F12, go to console tab, leave it open, then click "Refreh" button, is there any error msg in the console tab? Also, click "Refresh Civitai Helper" Button, is there any error msg in the console tab?
If there is any error msg, offer them to here.
I have out of memory errors and I get another error too:
I deleted the preview image and info and model from the gdrive folder. after that i get this error every time:
Applying xformers cross attention optimization. No Image data blocks found. Error loading embedding bad artist preview.png: Traceback (most recent call last): File "/content/gdrive/.shortcut-targets-by-id/1DiFC-YqwXU0EiVZqqevuNx2vsOHmJemo/sd/stable-diffusion-webui/modules/textual_inversion/textual_inversion.py", line 206, in load_from_dir self.load_from_file(fullfn, fn) File "/content/gdrive/.shortcut-targets-by-id/1DiFC-YqwXU0EiVZqqevuNx2vsOHmJemo/sd/stable-diffusion-webui/modules/textual_inversion/textual_inversion.py", line 155, in load_from_file name = data.get('name', name) AttributeError: 'NoneType' object has no attribute 'get'
But anyway, this extension is very helpful, thank you!
Read your error msg carefully, it is nothing to do with this extension. You removed "embedding bad artist preview.png" which is needed.
If you can not read an error msg, then do not mess up your files manually.
Also, it looks like a runtime error. Have you tried shutdown then relaunch your SD webui?
@butaixianran that was not an embedding. it is the preview image created by the extension for an embedding that i removed. somehow the extension wrote it as an embedding. that's what i think. i tried restarting many times. i still get the same error.
It is not an error from this extension, it already told you it is an error from xformers, which try to do some optimization for ti when SD's startup.
There is no such thing as "treat a png image as a ti" in this extension. A ti can be as a .png file, it is supported by xformers, that's why it is trying to find this image. It will know it is just an image not a ti after it read this file.
But You removed the real file and does not remove a shortcut. So it is always trying to check this file.
Again. Do not manually mess up cloud driver and colab unless you can read your error msg.
@butaixianran I mean other than manual management there is no other way to manage tls and loras. so thats what happened. do u have any suggestions, what to try?
I can only show you how to fix this issue.
You are using it on colab right? And you set a link from colab to your google drive, right?
Then you removed the real file in google drive, but didn't remove it in colab, right? So just remove that shortcut from colab.
@butaixianran Im using Lastben's fast stable diffusion from the github link. btw, I havent downloaded the notebbok to my drive. all other files are on my drive. I enter the path to my models in the notebook and run it. I really didnt understand where to delete shortcut in colab. which directory is that or do i need to edit a file?
oh man you did the fucking best job!!!
Why not use github URL to install?
You can, it is recommended, I'm not sure why you ask this.
how to edit the location of lora folder? Want to edit to root/extenstion/addation-network/....
SD webui has build-in support for lora, there is no need to use extension "addition-network" anymore.
You should move your lora back to SD webui's model/lora folder.
You can use the 'mklink' command in CMD with administrator privileges to link your any folder with the model folder of the webui
Do not try to sell mklink to people who don't familiar with it. It is complex and unnecessary for SD webui.
If you really need to change lora folder, you can use SD webui's command line: "--lora-dir".
Check SD webui's command line wiki:
https://github.com/AUTOMATIC1111/stable-diffusion-webui/wiki/Command-Line-Arguments-and-Settings
Learned something new, thanks a lot
How do you find the tab to check and download model/lora versions? It doesn't seem to be in my Civitai Helper tab.
Shutdown and Re-launch your SD webui after updating, just Reload UI won't work. As document already mentioned.
Also, if you are using an old version of SD webui, you need to update it.
nice
首先，你更新插件之后，按照要求重启SD webui了吗？只是Reload UI是无效的，文档说过了。
然后，Model Info is empty这是预览图的提示信息，不是错误。
下载模型提示的信息，你要往前看，前面Civitai给的报错是什么，导致你下载失败。你的插件能连上Civita吗？
Needs an option to skip downloading previews or only generating the info file on scan
The issue is the previews downloaded are appended with .preview which is only necessary for textual inversion embeddings, while hypernetworks, checkpoints, loras don't need it and saving a preview from a generated image does not come with the .preview in the file name so this ends up downloading a lot of additional images that don't get used since it doesn't see that the preview image (without '.preview') already exists
Read document, preview image part. You need to learn how extra network's preview image works.
I already have preview images for all the models (and they appear on the extra networks), I had originally used the modelname.preview.png naming format when manually saving these, but found that the button to save the generated image as preview created one that did not contain the .preview part of the name and this image took precedence over the other one, resulting in two images in the folder where the .preview.png file became useless. The only exception to this was the embeddings (textual inversions) where it has to be .preview
you can remove them by searching .preview.png
@butaixianran Yes, but the suggested option to skip downloading a preview image would be useful when updating info for new models vs doing that every time. I've edited the script locally to address this, but it'd be useful for people who were already managing preview images before
Failed to get model_info, check console log for detail 怎么办
都告诉你了在命令行界面查看详情，就点过去看一眼啊
另外，安装之后重新启动SD了吗，只是reload ui不起作用
Latest update broke the extension
No, it didn't. Like document said, re-launch SD, not just reload-ui. Read it.
Edit* It works, I just needed to use the 'Scan' button on the Civitai Helper tab, not just the 'Refresh Civitai Helper' button on the txt2img tab.
It worked great, then I updated (v1.5.3) and the 'Refresh Civitai Helper' button stopped working. I'm using the latest Automatic1111 (heads/master-0-gdfeee786), restarted it. There's nothing on the console when I press the button.
The tool still works on models that already have a settings file. I can't use the tool on new models.
What you mean stop working? those 4 icon buttons don't show on your model cards?
Then why are you also saying "extension still works on models that already have a settings file"? If those 4 buttons don't show on cards, then None of those models works.
Describe your issue right if you need answer.
Also, the latest version of SD webui is dfeee78, there is no such thing as master-0-gdfeee786. Are you using a modified fork?
The 4 icons work for images that were already refreshed before I updated Civitai Helper. The version was from git. I used a different command before. I'm a git noob. Anyways, the version from the last few days.
"git log --format="%H" -n 1
dfeee786f903e392dbef1519c7c246b9856ebab3"
I downloaded a bunch of pose extensions at the same time I updated Civitai Helper. Doing this broke Automatic1111. I deleted /venv and all of my extensions and started over. Everything works but the "Refresh Civitai Helper" button that fetches all of the preview images etc. Maybe it's a me problem.
refresh Civitai Helper button won't fetch any preview image. it just show 4 buttons on your extra network's model cards.
Do you mean scan model function don't work? Have you scanned new models after you download them?
You're right. I need to run the "Scan" button on the Civitai Tab, not just the refresh button on the txt2img tab. It works!
Works well! My only preference so far would be to have the option store the civitai.info and preview files in the extension subfolder, as now my models folder has 3x the files
I always get out of memory errors when reaching this model:
Skin Tight Tank Top - [CLOTHING][ANIME] | Stable Diffusion LORA | Civitai
I tried removing it, but still after going through some models, it gets stuck on another model and gets the same error.
for reference, im running colab and the ram usage goes over 12gb.
which function are you talking here, scanning or downloading?
@butaixianran when its scanning.
ok, since version 1.3, scanning already uses chunk to only read file with 4kb at a time. So, I need more detailed error msg about this issue.
Could you offer the full error msg of colab to here or github? You can open an issue about this. A sceenshot of error msg also works.
Also, have you re-launched SD webui after installing or updating? Just reload-ui won't work.
ok, updated a v1.5.5 to github, to make this sha256 function even better. Check it out, also re-launch SD webui after updating.
If this update can not solve your issue, then there is nothing I can do. You need to search your error msg in google, in that case.
@butaixianran Oh that's great. I dont get an error log anymore. it just stops and shows error in the webui. I have to restart to use again. anyway, I'll try the latest update and see what i can do. praying lol
更新了版本还是无法显示那个四个小按钮
* 你的SDwebui版本号贴过来对照下，SD版本号在页面最下面，有个叫 commit: xxxx 就是版本号，那个版本号是个链接，整个链接地址贴过来。
* 你点"Refresh Civitai Helper"按钮了吗？
* 你是小图模式吗，如果是小图模式，插件tab最下面，有个小图模式的开关选项，打开才显示。
在github中文文档最后加入了常见问题部分，包含了这个的说明。查看文档更新
@butaixianran
python: 3.10.8 • torch: 1.13.1+cu117 • xformers: 0.0.16rc425 • gradio: 3.16.2 • commit: 0cc0ee1b • checkpoint: f90d39fdda
已经照着教程做了，还是找不到那四个小按钮。。。
升级sd webui ，看不到就说明没升级。中文文档最后常见问题部分有这个的说明。
更新了中文文档，查看最后常见问题部分。如果你用的国内汉化版，上面的信息应该有帮助。
感谢您的工作，来自美国。你是这个社区的英雄，你的奉献和热情是显而易见的。祝你成功和幸福。
点了这四个小图标没有反应 控制台显示：Load model info of /breastinclassBetter_v141.safetensors in lora
你点的没反应的这个模型，有从civitai取到模型信息吗？模型信息文件是空的吗？
@butaixianran 执行了但提示找不到模型信息文件,无法获取 lora
那就是模型信息是空的，civitai没有这个模型的话，你本地没有模型信息，那自然就点按钮没有反应。
关于为什么在civitai上找不到这个模型，阅读文档的 SHA256部分。
谢谢，已解决 但发生一个新的问题 获取封面按键没反应 其他三个有反应
控制器提示：There is no image in gallery to save as a preview.
那不是获取封面按钮，那是把当前生成的图片保存为预览图。是SD webui自带的功能。
@butaixianran 谢谢 我看图标换了 就把他理解成获取封面了
You are a hero!
Great idea. Thanks!
hello dalao,
I am at latest version, why is there logs like "Civitai Link: Presence update: SD: xxx, Clients: x" repeatly print in the backend termainal?
You are not even using this extension.
this extension's name is Civitai Helper, not Civitai Link.
There is no such log in this extension
webui - settings - Civitai - uncheck 'Show Civitai Link events in the console'
Can you help me with the error? it shouldn't be from the extension, but idk what to do :(
Every time I try to run SD in colab, i get this:
reading lora /content/gdrive/MyDrive/sd/stable-diffusion-webui/models/Lora/yamamuraSadakoTheRing_v3.safetensors: OSError Traceback (most recent call last): File "/content/gdrive/MyDrive/sd/stable-diffusion-webui/extensions-builtin/Lora/lora.py", line 53, in init File "/content/gdrive/MyDrive/sd/stable-diffusion-webui/modules/sd_models.py", line 216, in read_metadata_from_safetensors OSError: [Errno 107] Transport endpoint is not connected: '/content/gdrive/MyDrive/sd/stable-diffusion-webui/models/Lora/yamamuraSadakoTheRing_v3.safetensors' reading lora /content/gdrive/MyDrive/sd/stable-diffusion-webui/models/Lora/youreUNDERARRESTLoha_v10.safetensors: OSError Traceback (most recent call last): File "/content/gdrive/MyDrive/sd/stable-diffusion-webui/extensions-builtin/Lora/lora.py", line 53, in init File "/content/gdrive/MyDrive/sd/stable-diffusion-webui/modules/sd_models.py", line 216, in read_metadata_from_safetensors OSError: [Errno 107] Transport endpoint is not connected: '/content/gdrive/MyDrive/sd/stable-diffusion-webui/models/Lora/youreUNDERARRESTLoha_v10.safetensors' Error loading script: lora_script.py Traceback (most recent call last): File "/content/gdrive/MyDrive/sd/stable-diffusion-webui/modules/scripts.py", line 248, in load_scripts File "/content/gdrive/MyDrive/sd/stable-diffusion-webui/modules/script_loading.py", line 11, in load_module File "<frozen importlib._bootstrap_external>", line 850, in exec_module File "<frozen importlib._bootstrap>", line 228, in callwith_frames_removed File "/content/gdrive/MyDrive/sd/stable-diffusion-webui/extensions-builtin/Lora/scripts/lora_script.py", line 5, in <module> File "<frozen importlib._bootstrap>", line 1007, in findand_load File "<frozen importlib._bootstrap>", line 982, in findand_load_unlocked File "<frozen importlib._bootstrap>", line 925, in findspec File "<frozen importlib._bootstrap_external>", line 1423, in find_spec File "<frozen importlib._bootstrap_external>", line 1395, in getspec File "<frozen importlib._bootstrap_external>", line 1526, in find_spec File "<frozen importlib._bootstrap_external>", line 1569, in fillcache OSError: [Errno 107] Transport endpoint is not connected: '/content/gdrive/MyDrive/sd/stable-diffusion-webui/extensions-builtin/Lora' Error loading script: scunet_model.py Traceback (most recent call last): File "/content/gdrive/MyDrive/sd/stable-diffusion-webui/modules/scripts.py", line 248, in load_scripts File "/content/gdrive/MyDrive/sd/stable-diffusion-webui/modules/script_loading.py", line 11, in load_module File "<frozen importlib._bootstrap_external>", line 846, in exec_module File "<frozen importlib._bootstrap_external>", line 982, in get_code File "<frozen importlib._bootstrap_external>", line 1039, in get_data OSError: [Errno 107] Transport endpoint is not connected: '/content/gdrive/MyDrive/sd/stable-diffusion-webui/extensions-builtin/ScuNET/scripts/scunet_model.py'
The list of errors like this goes on after this.
But, if I remove the Lora folder before running SD, I dont get the error. But I can't use my Loras then.
Any help is appreciated, thanks.
hi, this is a corlab error, there is nothing to do with this extension. Search "Transport endpoint is not connected corlab" in Google to learn it.
I can only support this extension, I have no time to teach people how to use corlab, learn it from Google.
@butaixianran I tried a few things and i think i figured out the problem:
Automatic ui has trouble loading many loras at once. i have 200 loras and their related files, when i only move half of them first, and the rest after loading, I don't have issues. Might get fixed soon i guess.
大佬求助，之前用的好好的（只有扫描功能时），今天更新了新版civitaihelper后四个按钮就不显示了
重启sdweb ui了吗，不光是reloadui。
sd webui 升级到最新版了吗。
如果使用了双语汉化插件，换成普通汉化插件。
@butaixianran 请问普通汉化插件指的是哪种，换了好几个都不显示这个4个按钮了
@92811jason865 如果关掉汉化插件还不显示，那就是你的SD webui太旧了，要升级。
如果你是昨天才升级的SD webui那就是SD webui界面突然大幅更新导致的。需要去github升级本插件的最新版v1.6
大佬，c站被墙了 你的插件也用不了了怎么办
我也是
挂全局代理
@butaixianran 挂了全局代理也不行😭
@242424 挂了代理就一定能正常。今天的话，是因为civitai网站挂了。它挂一次8小时，差不多2~3周挂一次。网站恢复了再试。
已支持代理，在github查看新版。用法看文档。
This could be integrated with Dynamic Prompts.
https://github.com/adieyal/sd-dynamic-prompts
Scan lora to create wildcard files based on tags.
Example: all lora with "character" tag would be put in a character.txt file alphabetically like
<A_character:1>, A_char_trigger
<B_character:1>, B_char_trigger
Then they could be used with the '__character__' wildcard. Same for '__style__'
The main issue I see is that tagging on civitai isn't good and relies on the author to tag accurately. If civitai implemented tag weights/voting it would be better.
今天stable diffusion更新了，你的插件也重新安装了最新版本，现在lora旁边的 插件按钮也消失了，但是页面最上方的助手选项有的，就是4个按钮消失了，希望大神更新下
更新最新版v1.6
我扫描不了本地文件了，用了梯子也没用。。。一直time out，请问怎么解决